Relationship between Retinopathy of Prematurity (ROP) and Bronchopulmonary Dysplasia (BPD): A Retrospective-Analytical Study

Introduction

Retinopathy of prematurity (ROP) is a developing retina disease. More premature neonates are predisposed to severe stages of ROP [1,2]. Despite it being the leading cause of infantile blindness all over the world, the disease is avoidable [3,4]. There has been an increase in global incidence [3] and severity [4] of ROP. Consequently, ROP is a significant disease that deserves further concern.  

Bronchopulmonary dysplasia (BPD) is one of the commonest complication of prematurity with a rising global incidence. There is an inverse proportionality between the incidence of BPD on one hand,   birth weight and gestational age on the other [5,6]. BPD and ROP share common risk factors of inflammation and oxidative stress exposure [7–11]. Moreover, both diseases have a genetic background [12,13]. Results from a previous study conducted at our center established a relationship between the incidences of ROP among BPD patients [14]. Therefore, it was decided that this connection should be investigated further. 

The aim of this study is to research the ROP prevalence and severity among preterms diagnosed with BPD and to examine the shared risk factors between ROP and BPD

Methods

This case-control study consisted of 44 preterm infants (cases) of less than or equal to 32 weeks gestation who had developed BPD and were admitted to the Neonatal Intensive Care Unit (NICU) of Aburreesh El-Mounira, Cairo University Children Hospital between January 1, 2019, and December 31, 2020.

In addition, 62 gestational age-matched controls who were admitted during the same period were also included in the study. Infants who died before being diagnosed with BPD and newborns with congenital anomalies were excluded. In all preterms, the presence and staging of ROP in both eyes were documented retrospectively. 

Risk factors associated with the development of BPD and ROP were documented in accordance with medical records, and the shared risk factors were identified. The data analyzed included gestational age  in weeks, birth weight in grams, sex, duration of admission in days, early onset sepsis (EOS), late onset sepsis (LOS) and premature rupture of membranes (PROM), patent ductus arteriosus (PDA), intraventricular hemorrhage (IVH), necrotizing enterocolitis (NEC), duration of mechanical ventilation, CPAP, oxygen blender, head box and incubator oxygen. In addition, the duration of oxygen in days, duration of TPN, caffeine citrate, and inotropes together with inhaled steroids were documented. Moreover, surfactant therapy, the number of blood transfusions and infectious episodes were also documented. 

BPD diagnosis and severity were determined by using BPD severity-based criteria proposed by the National Institute of Child Health and Human Development (NICHD) [15]. At the time of diagnosis, infants who did not require oxygen were considered as mild BPD cases. Moderate BPD was considered for cases requiring less than 30% oxygen, and severe BPD was considered for cases needing positive pressure and/or oxygen support ≥30% [15]. 

ROP was diagnosed by a neonatologist and confirmed by an ophthalmologist according to the American Academy of Ophthalmology definition [16]. Based on the diagnosis and staging of ROP, all preterm infants were classified into the following groups: absent, low grade (stage 1 or 2) severe ROP (more than stage 2 with or without plus disease).

Sepsis was considered following a positive blood culture result and evident symptoms. Early-onset sepsis is sepsis occurring before seven days and late-onset sepsis is sepsis occurring seven days or more after birth. NEC was considered according to symptoms, radiographic findings, and the staging criteria modified by Bell et al. [17]. Moreover, PDA was considered by assessing symptoms and echocardiography results. Furthermore, Intracranial hemorrhage (ICH) was identified by performing a brain ultrasound [18–20].

Ethical considerations

This study’s protocol was accepted by the Ethical Committee of the Faculty of Medicine, Cairo University, and corresponded with the provision of the Declaration of Helsinki in 1964 and its later amendments or comparable ethical standards. Informed signed consent was collected from the parents of each preterm before their inclusion in the study

Statistical analysis

Data analysis was studied using the IBM SPSS Statistics program version 21. Quantitative variables were displayed as mean, standard deviation (SD), median, and interquartile range (IQR), while qualitative variables were defined as numerical values and percentage. Comparison between quantitative variables was done using independent T test if variables were normally distributed or using mannwhitney test in case of non-normally distributed variables A chi-square test was employed to compare qualitative variables and the Pearson correlation coefficient (PCC) was used to test the linear relationship between variables. Post hock test with Bonferroni adjustment was used in comparing qualitative variables of more than two categories. A p-value less than or equal to 0.05 was considered significant. In contrast, a p-value less than or equal to 0.01 was deemed highly significant. Multiple stepwise backward logistic regressions were conducted to detect the significant predictors for BPD and ROP.

NB Stepwise regression was conducted through several steps. In each step, the insignificant variables were excluded from the model. Accordingly, those reported in the last step were only the significant variables.

Results

During the two-year study period, 106 premature infants with less than or equal to 32 weeks gestation were included in the study. There were 44 cases and 62 gestational age-matched controls as shown in figure (1). The BPD cases when compared to controls: (a) average gestational age was 30.77±1.41 (mean ± standard deviation); (b) average birth weight in grams was 1262.56±216.7; (c) average admission period in days was 59.25±24.31; median48(40.5:79) (d) average duration of oxygen in days was 44.68±20.3; median 38(32:47) (e) average duration of mechanical ventilation, CPAP, oxygen blender, head box, and incubator oxygen were 24.18 ± 22.67 median 13(7:35), 15.11 ± 9.2 , median 14.5 (11:18),  8.18±4.43, median 8.5 (6:12), 3.09 ± 2.24, median 3(1:5), 2.91 ± 2.15, median 3(1.5:4), respectively. 

Furthermore, the average duration of TPN in days was 24.73±15.17 median 22.5(12:40); (f) average duration of caffeine citrate was 42.68±22.84; median37.5(23.5:54) (g) average intake of inotropes duration in days was 14.11±10.02; median16(11:24) (h) average inhaled steroids duration in days was 27.23±17.9 median 23(14:38); (i) the average BPD regimen duration in days was 14.7±9.8 Finally, (j)the average infectious episodes were 3.5±1.66 ,median 3(2:5); and the average intake of PRBCs was 2.67±1.8 median 2(1:4) times as shown in Table 1(A)

BPD cases were significant to birth weight in grams and intake of PRBCs with a p value of 0.004 for both. BPD cases were highly significant (p <0.001) to admission duration, oxygenation duration, durations of mechanical ventilation, CPAP, oxygen blender, head box and incubator oxygen. BPD cases were highly significant too to durations of total parenteral nutrition and medications  administration in the form of caffeine citrate, inotropes, and inhaled steroids(P<0.001). BPD cases correlated highly to plasma intake and infectious episodes (p<0.001) as shown in Table 1(A).

BPD cases were consistent with administration of TPN regardless of duration (p0.029) and were highly significant (p<0.001) to administration of mechanical ventilation, CPAP, inhaled steroids and inotropic support regardless of duration too. Regarding the clinical course of the preterms involved, surfactant administration, presence of PDA, development of NEC, IVH, and pneumothorax correlated significantly to BPD (p<0.001)   as shown in Table 1(B). 

ROP cases were correlated with administration of oxygen blender (p 0.003) head box (p0.005), incubator oxygen (p 0.012), inotropic support (p 0.01), surfactant (p0.004), and plasma transfusion (p 0.004). Presence of PDA (p0.036) and development of IVH (0.042) correlated too to ROP cases. They highly correlated with administration of mechanical ventilation, CPAP, inotropic support, inhaled steroids, PRBCs transfusion and BPD regimen (p<0.001). Presence of PDA and development of LOS highly correlated with ROP (p<0.001) as shown in Table 2(A). Distribution of ROP cases is displayed in figure (2)

ROP cases showed statistical significance to birth weight (p0.025) and duration of incubator oxygen (p0.019). They were highly significant to durations of ; admission, oxygenation, oxygen blender and CPAP administration (p0.001). They were also highly significant to durations of; mechanical ventilation (p0.004), head box (p0.002), caffeine citrate (p0.005) and total parenteral nutrition. Regarding infectious episodes ROP cases were highly significant to them (p0.003) as shown in Table 2(B)

Both BPD and ROP corresponded with the durations of the administration of CPAP, oxygen blender, head box, incubator oxygen, mechanical ventilation, duration of admission, oxygenation, caffeine citrate, TPN, administration and duration of inhaled steroids, inotropic support, surfactant administration, PDA, ICH, PRBCs and plasma transfusion, LOS and infectious episodes as shown in Table (1) and Table (2). Severe cases of ROP occurred in BPD cases, and this connection extended to varying grades of both diseases as shown in Table (3) and Figure (3). The use of inotropic support was the most predisposing factor to BPD as shown in the multiple regression analysis in Table (4) The model showed significance for BPD at X2 (85.71) and p-value

Discussion

Our study revealed a link between the incidence of ROP in preterm infants with the incidence of BPD. This is consistent with the findings presented by Guven et al., [21], Kornacka et al., [22], Higgins et al., [23] and Leviton et al., [24]. The presence of BPD necessitates oxygen therapy that in turn results in increased incidences of ROP due to immaturity of the antioxidant systems together with the use of steroids. Unfortunately, targeting oxygen saturation below 90% is correlated with a higher risk of death [25]. Therefore, neonatologists are obliged to target higher levels, especially ROP is treatable.

Consistent with the study results, severe cases of ROP occurred in BPD cases, and this relationship extended to varying grades of both diseases. This finding corresponded with the results of Krzysztofowicz et al., and many other studies [26-29].

According to Shah et al., [28] and Allegaert et al., [29], it was concluded that extended duration of mechanical ventilation and/or nasal continuous airway pressure (nCPAP) played a significant role in developing severe forms of ROP. Yang et al., [30] and Seiberth et al., [31] concluded that mechanical ventilation increases the severity of ROP. Previous studies [32–34] concluded that developing BPD and ROP is affected by free radicles generated secondary to oxygen excess and increased partial oxygen pressure during mechanical ventilation and oxygenation. This primarily applied to this study, where both BPD and ROP were associated with durations of CPAP, oxygen blender, head box, incubator oxygen and mechanical ventilation, admission duration and oxygenation. Another explanation is that it may be related to the fact that our center is a tertiary center. The nature of the cases referred to us from other hospitals under weak transport facilities exposes patients to alternate episodes of hypoxia, hyperoxia, hypothermia and hypotension, and this adds to the unstable preterm infants from the start. Mohamed et al. [35] mentioned that alternating hypoxia and hyperoxia were more dangerous than hypoxia or hyperoxia alone. As for BPD, alternating hypoxia and hyperoxia precipitates oxidative stress causing more cellular and pulmonary inflammation. In case of ROP, after weaning from oxygen the retina becomes relatively hypoxic inducing angiogenic factors causing vasoproliferation of retinal vessels and worsening of the grades of ROP[35],[36].

In addition, Yue et al., [37] stated that smaller gestational age, lower blood pressures, decreased Apgar score at 5 min, increased respiratory rate, patent ductus arteriosus, increased C-reactive protein levels, all of which were classified as significant risk factors of adopting mechanical ventilation in preterms. These factors were shared by the vast majority of preterms included in our study, many of whom ended up in mechanical ventilation, which may impose a clue to ending up in both diseases.

Both BPD and ROP were associated with the administration and duration of caffeine citrate and inhaled steroids [14], which is typically the routine for younger preterms that require a longer duration of admission and respiratory support. Extended caffeine and inhaled steroids duration reflected a younger preterm. In an animal study, Poon et al., [38] discovered that hyperoxia could cause injury to the brain, lungs and retinas, adding that more significant lung injuries were correlated with more significant retinal and brain injuries.

Both BPD and ROP share common pathogenesis [39-45], and both have dysfunction of angiogenic signaling pathways [39-50], disruption of which causes altered development and vasculature of both retinal and pulmonary vessels.

According to Podraza et al., [27], Del Vecchio et al., [51], Fortes et al.,[52], Northway et al.,[53], Englert et al., [54], and Gomaa & Abdelkhalik [55], frequent blood transfusions were found to increase the risk of both BPD and ROP, and these results were confirmed in this study. Oxidative injury caused by a blood transfusion increases non-transferrin bound iron. Furthermore, the inflammatory mediators present in stored blood products may explain the association between ROP and BPD on one hand and plasma transfusion on the other. However, another explanation may be due to the use of plasma and PRBCs as colloids during resuscitation of unstable preterms with hypotension. Blood transfusions increase the risk of ROP by two mechanisms: (a) by an increase in retinal oxygen supply and (b) by an increase in oxygen free radicals through free iron overload.

In the multiple regression analysis, PDA was found to be a risk factor for BPD and this coincided with Kim et al., Ding et al., and Abuelhamd et al., [56-58]. Terrin et al., [59] stated that hemodynamically unstable PDA is a predictor for BPD, IVH, ROP, increased mechanical ventilation duration and hypotension and added that together with birth weight, gestational age, and prenatal steroids, there is a significant association between hypotension and hemodynamically unstable PDA that may partly explain the higher incidence of inotropic support and duration in this study. El Sayed and Fraser [60] stated that hemodynamic instability caused by PDA may predispose to a variety of organ injuries causing clinical complications.  Additionally, ROP was an independent risk factor in our study. The intake of inotropes and the duration of admission in days [58]were independent risk factors for BPD. The intake of inotropes was the most significant predictor denoting that preterms who received inotropes were five times more likely to develop BPD than those who did not [61]. Whether this is solely related to inotropes, hemodynamic instability and hypotension [56] associated with prematurity or with a hemodynamically significant PDA needs further investigation. Regarding ROP, mechanical ventilation, intake of oxygen and the duration of admission in days were independent risk factors in the multiple regression analysis, with mechanical ventilation being the most significant factor.

Ventilated preterms were five times more likely to develop ROP than those who were not ventilated. This was in accordance with a study conducted by Madhu et al., [62].

Limitations: The limitation of this retrospective study is the improper recording of data that would have helped achieve better results. Moreover, better outcome would have have been achieved with more number of patients

Conclusions

BPD and ROP share common risk factors, and there is a connection between them as regards the varying grades of severity. Though, there is some evidence that hemodynamic instability, longer inotropic support, hemodynamically significant PDA, prolonged ventilation act as cofactors between both diseases.  

Recommendations

It is of the utmost importance to limit duration of admission, oxygenation, ventilation and TPN. Emphasis should be made on judicious use of inotropes and minimizing their duration to the least possible. Beginning with the end in mind, should be the slogan when ventilating a neonate or administering inotropes. Though, we have to admit a compromise is difficult owing to the link between inotropes, PRBCs and plasma transfusions on one hand together with ROP and BPD on the other. Further research is needed to correlate between ROP and BPD and the aforementioned results

Abbreviations

BPD:  Bronchopulmonary Dysplasia

ROP: Retinopathy of Prematurity

EOS: Early-Onset Sepsis

LOS: Late-Onset Sepsis

NEC: Necrotizing Enterocolitis

PDA: Patent Ductus Arteriosus

ICH: Intracranial Hemorrhage

IUGR: Intra-Uterine Growth Restriction

TPN: Total Parenteral  Nutrition

GA: Gestational Age

BW: Birth Weight

RBC: Red Blood Cells

PRBC: Packed Red Blood Cells

NICU: Neonatal Intensive Care Unit

CPAP: Continuous Positive Airway Pressure

MV: Mechanical Ventilation

PROM: Premature Rupture of membranes